Method and device for predicting extracorporeal shock wave lithotripsy reflecting sinogram and / or inferred attenuation information of shock waves

Abstract

A method for predicting an extracorporeal shock wave lithotripsy according to an embodiment may comprise the steps of: obtaining a first image including target stone information by inputting a computed tomography image to a first artificial neural network model; obtaining a second image including reflection information of shock waves on the basis of the computed tomography image; and obtaining a sinogram including attenuation information of the shock waves and the reflection information on the basis of the first image and the second image.

Description

Method and device for predicting extracorporeal shock wave lithotripsy reflecting sinogram and / or attenuation information of inferred shock waves

[0001] The following examples relate to a system for predicting the success of extracorporeal shock wave lithotripsy. More specifically, the examples relate to a method, device, and system for predicting the success of extracorporeal shock wave lithotripsy by reflecting sinogram and / or inferred shock wave attenuation information.

[0002]

[0003] Extracorporeal shock wave lithotripsy (ESWL) is a treatment method that crushes and treats kidney stones by applying shock waves generated outside the body to them. Clinical efforts have attempted to predict the success of ESWL by analyzing the characteristics of stone shadows on computed tomography (CT), but this has limited predictive power.

[0004]

[0005] A method for predicting extracorporeal shock wave lithotripsy according to one embodiment may include the steps of: inputting a computed tomography image into a first artificial neural network model to obtain a first image including target stone information; obtaining a second image including reflection information of a shock wave based on the computed tomography image; and obtaining a sinogram including attenuation information of the shock wave and the reflection information based on the first image and the second image.

[0006] The method for predicting extracorporeal shock wave lithotripsy according to one embodiment may further include a step of acquiring meta information of a patient corresponding to the computed tomography image.

[0007] A method for predicting extracorporeal shock wave lithotripsy according to one embodiment may further include a step of predicting the success or failure of extracorporeal shock wave lithotripsy by inputting the sinogram, the first image, the second image, and the patient's meta information into a second artificial neural network model.

[0008] The second artificial neural network model includes a prediction module composed of a plurality of prediction blocks, and the prediction module can determine an output of the prediction module by combining the outputs of each of the plurality of prediction blocks.

[0009] The step of acquiring the first image may include a step of inputting the computed tomography image into the first artificial neural network model to acquire a target stone location; and a step of modifying the computed tomography image based on the target stone location to generate the first image.

[0010] The step of acquiring the first image may include a step of inputting the computed tomography image into the first artificial neural network model to acquire texture information of the target stone.

[0011] The step of acquiring the second image may include a step of extracting a boundary region of the computed tomography image; and a step of converting the reflection information of the boundary region into the attenuation information.

[0012] A prediction device according to one embodiment may include a processor that inputs a computed tomography image into a first artificial neural network model to obtain a first image including target stone information, obtains a second image including reflection information of a shock wave based on the computed tomography image, and obtains a sinogram including attenuation information of the shock wave and the reflection information based on the first image and the second image.

[0013] The above processor can obtain meta information of the patient corresponding to the above computed tomography image.

[0014] The above processor can input the sinogram, the first image, the second image, and the patient's meta information into a second artificial neural network model to predict the success or failure of extracorporeal shock wave lithotripsy.

[0015] The second artificial neural network model includes a prediction module composed of a plurality of prediction blocks, and the prediction module can determine an output of the prediction module by combining the outputs of each of the plurality of prediction blocks.

[0016] The processor can input the computed tomography image into the first artificial neural network model to obtain a target stone location, and modify the computed tomography image based on the target stone location to generate the first image.

[0017] The above processor can input the computed tomography image into the first artificial neural network model to obtain texture information of the target stone.

[0018] The above processor can extract a boundary region of the computed tomography image and convert the reflection information of the boundary region into the attenuation information.

[0019]

[0020] Figure 1 is a diagram illustrating an example of a mathematical model that simulates the behavior of a biological neuron.

[0021] Figure 2 is a diagram illustrating an example of a neural network.

[0022] Figure 3 is an exemplary diagram of a prediction device configuration according to one embodiment.

[0023] FIG. 4 is a drawing for explaining a method for predicting extracorporeal shock wave lithotripsy according to one embodiment.

[0024] FIG. 5 is a diagram illustrating an example of a method for predicting extracorporeal shock wave lithotripsy according to one embodiment.

[0025] FIG. 6 is a drawing for explaining a sinogram including attenuation information of a shock wave according to one embodiment.

[0026] FIG. 7 is a drawing for explaining a modified sinogram according to one embodiment.

[0027] FIG. 8 is a drawing for explaining a method for obtaining a second image according to one embodiment.

[0028] FIG. 9 is a diagram illustrating a structure of a second artificial neural network according to one embodiment.

[0029]

[0030] The specific structural or functional descriptions disclosed in this specification are merely illustrative for the purpose of explaining embodiments according to technical concepts, and the actually implemented form may have various different appearances and is not limited to the embodiments described in this specification.

[0031] While terms like "first" and "second" may be used to describe various components, these terms should be understood solely to distinguish one component from another. For example, a "first" component may be referred to as a "second" component, and similarly, a "second" component may also be referred to as a "first" component.

[0032] When a component is referred to as being "connected" or "connected" to another component, it should be understood that it may be directly connected or connected to that other component, but that there may be other components in between. Conversely, when a component is referred to as being "directly connected" or "directly connected" to another component, it should be understood that there are no other components in between. Expressions that describe relationships between components, such as "between" and "immediately between," or "adjacent to" and "directly adjacent to," should be interpreted similarly.

[0033] Singular expressions include plural expressions unless the context clearly dictates otherwise. In this specification, the terms "comprises" or "has" should be understood to indicate the presence of a feature, number, step, operation, component, part, or combination thereof, but not to exclude the possibility of the presence or addition of one or more other features, numbers, steps, operations, components, parts, or combinations thereof.

[0034] Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by a person of ordinary skill in the art. Terms defined in commonly used dictionaries should be interpreted to have a meaning consistent with their meaning in the context of the relevant technology, and will not be interpreted in an idealized or overly formal sense unless explicitly defined herein.

[0035] The embodiments can be implemented in various forms of products, such as personal computers, laptop computers, tablet computers, smartphones, televisions, smart home appliances, intelligent vehicles, kiosks, and wearable devices. The embodiments are described in detail below with reference to the attached drawings. Like reference numerals in each drawing represent like elements.

[0036] Computed tomography is a tomography technique that reconstructs cross-sectional images using X-ray images taken from various directions. It can be referred to as computed tomography scan (CT scan), X-ray computed tomography (X-ray CT), or computed axial tomography. In computed tomography, an X-ray generator and an X-ray detector rotate in pairs around the subject to capture X-ray images from all directions. The original image obtained in this way is called a sinogram.

[0037] Afterwards, the sinogram can be restored by calculating the inverse Radon transform on a computer to obtain a cross-sectional image. In this cross-sectional image, each pixel represents the attenuation rate of each part. The cross-sectional image obtained in this way can be converted into a complete 3D image through geometric processing. In the 3D image created in this way, each pixel is called a voxel rather than a pixel. When discussing attenuation rate, Hounsfield units (HU), a relative unit compared to the attenuation rate of water, are used.

[0038] Extracorporeal shock wave lithotripsy (ESWL) is a treatment method that crushes and treats kidney stones by applying shock waves generated outside the body to them. As mentioned above, existing techniques attempt to predict the success of ESWL by analyzing the characteristics of stone shadows on computed tomography (CT), but these techniques have limited predictive power.

[0039] A method for predicting extracorporeal shock wave lithotripsy according to one embodiment relates to a deep learning-based prediction method for predicting the treatment success of extracorporeal shock wave lithotripsy by combining the results of texture analysis of stones and human image analysis based on the shadow intensity (Hounsfield unit, HU) of computed tomography with clinical data. Before describing the method for predicting extracorporeal shock wave lithotripsy according to one embodiment, a neural network is described with reference to FIGS. 1 and 2.

[0040] Figure 1 is a diagram illustrating an example of a mathematical model that simulates the behavior of a biological neuron.

[0041] Referring to Fig. 1, various attempts are being made to develop a computing device that can efficiently process a large amount of information by simulating biological neurons or biological neural networks formed by interconnected biological neurons. The operation of a biological neuron can be simulated by a mathematical model (11). The mathematical model (11) may be an example of a neuromorphic operation executed by a hardware computational device or a hardware computational processor. The mathematical model (11) may include a multiplication operation that multiplies information from a plurality of neurons by a synaptic weight, an addition operation (Σ) for values ​​(w0x0, w1x1, w2x2) multiplied by the synaptic weight, and an operation that applies a characteristic function (b) and an activation function (f) to the result of the addition operation. By executing the neuromorphic operation, a neuromorphic operation result can be provided. Here, values ​​such as x0, x1, x2, ... correspond to axon values, and values ​​such as w0, w1, w2, ... correspond to synaptic weights.

[0042] Figure 2 is a diagram illustrating an example of a neural network.

[0043] Referring to FIG. 2, an example of an artificial neural network, i.e., a neural network (20) that mimics a neural network formed by interconnecting biological neurons, is illustrated. The neural network (20) may be an example of a deep neural network (DNN). For convenience of explanation, the neural network (20) is illustrated as including two hidden layers, but may include a variety of hidden layers. In addition, although the neural network (20) in FIG. 2 is illustrated as including a separate input layer (21) for receiving input data, the input data may be directly input to the hidden layer.

[0044] In a neural network (20), artificial nodes in layers other than the output layer can be connected to artificial nodes in the next layer via links for transmitting output signals. Through these links, the output of an activation function regarding the weighted inputs of artificial nodes included in the previous layer can be input to the artificial node. The weighted input is the input (node ​​value) of the artificial node multiplied by a weight, where the input corresponds to axon values ​​and the weight corresponds to synaptic weights. The weight may be referred to as a parameter of the neural network (20). The activation function may include a sigmoid, a hyperbolic tangent (tanh), and a rectified linear unit (ReLU), and nonlinearity may be formed in the neural network (20) by the activation function.

[0045] The output of any node (22) included in this neural network (20) can be expressed as in the following mathematical expression 1.

[0046]

[0047] Mathematical expression 1 can represent the output value yi of the ith node (22) for m input values ​​in any layer. xj can represent the output value of the jth node of the previous layer, and wj,i can represent the output value of the jth node and the weight applied to the ith node (22) of the current layer. f() can represent an activation function. As shown in Mathematical expression 1, for the activation function, the accumulated result of the multiplication of the input value xj and the weight wj,i can be used. In other words, the operation of multiplying and adding the appropriate input value xj and the weight wj,i (MAC operation) can be repeated at a desired point in time.

[0048] Figure 3 is an exemplary diagram of a prediction device configuration according to one embodiment.

[0049] Referring to FIG. 3, the prediction device (300) may include a processor (301), a memory (303), and a sensor (305).

[0050] The processor (301) can input a computed tomography image into a first artificial neural network model to obtain a first image including target stone information, obtain a second image including reflection information of a shock wave based on the computed tomography image, and obtain a sinogram including attenuation information and reflection information of the shock wave based on the first image and the second image.

[0051] The processor (301) obtains meta information of a patient corresponding to a computed tomography image, and inputs the sinogram, the first image, the second image, and the meta information of the patient into a second artificial neural network model, thereby predicting the success or failure of extracorporeal shock wave lithotripsy.

[0052] The processor (301) can input a computed tomography image into a first artificial neural network model to obtain a target stone location, and modify the computed tomography image based on the target stone location to generate a first image.

[0053] The processor (301) can input a computed tomography image into a first artificial neural network model to obtain texture information of a target stone. The processor (301) can extract a boundary region of the computed tomography image and convert reflection information of the boundary region into attenuation information. The specific operation method of the processor (301) is described in detail below with reference to FIGS. 4 to 9.

[0054] The memory (303) according to one embodiment may be a volatile memory or a non-volatile memory, and the memory (303) may store data necessary to perform an operation.

[0055] A sensor (305) according to one embodiment can receive a computed tomography image.

[0056] The prediction device (300) according to one embodiment may further include other components not shown. For example, the prediction device (300) may further include an input / output interface including a communication module and an input device and an output device as means for interfacing with the communication module. Furthermore, for example, the prediction device (300) may further include other components such as a transceiver, various sensors, a database, etc.

[0057] FIG. 4 is a drawing for explaining a method for predicting extracorporeal shock wave lithotripsy according to one embodiment.

[0058] Referring to FIG. 4, a prediction device (for example, a prediction device (300) of FIG. 3) can input an input image (410) into an artificial neural network model to obtain a deep learning image (420). The input image according to one embodiment may include a computed tomography image. The computed tomography image may be a cross-sectional image obtained by restoring a sinogram through an inverse Radon transform calculation. Alternatively, the computed tomography image may be an image obtained by converting a cross-sectional image into a three-dimensional image. The artificial neural network model may be a model trained to receive an input image (410) and output a deep learning image (420) including target stone information. The artificial neural network model according to one embodiment may automatically classify a target kidney stone and extract the location of the target kidney stone.

[0059] A prediction device according to one embodiment can generate a modified sinogram (430) based on a deep learning image (420). The modified sinogram (430) is a sinogram different from the sinogram used to generate the input image (410), and can include shock wave attenuation information and reflection information.

[0060] According to one embodiment, a prediction device can predict the success of extracorporeal shock wave lithotripsy based on a deep learning image (420) and a modified sinogram (430). The prediction device can improve the accuracy of predicting the success of extracorporeal shock wave lithotripsy by utilizing not only the deep learning image (420) and the modified sinogram (430), but also patient metadata. Below, an example of a method for predicting the success of extracorporeal shock wave lithotripsy is described with reference to FIG. 5 .

[0061] FIG. 5 is a diagram illustrating an example of a method for predicting extracorporeal shock wave lithotripsy according to one embodiment.

[0062] Referring to FIG. 5, a prediction device (e.g., a prediction device (300) of FIG. 3) can input a computed tomography image (510) into a first artificial neural network model to obtain a first image (520) (e.g., a deep learning image (420) of FIG. 4).

[0063] The prediction device can obtain texture information of a target stone by inputting a computed tomography image (510) into a first artificial neural network model. The prediction device can obtain texture information including the three-dimensional size of the stone, average brightness information of the stone, standard deviation of the brightness information of the stone, location information of the stone, etc., which can identify features in the computed tomography image (510).

[0064] The prediction device can input a computed tomography image (510) into a first artificial neural network model to obtain a target stone location, and modify the computed tomography image (510) based on the target stone location to generate a first image (520). A specific method of modifying the computed tomography image (510) to generate the first image (520) is described with reference to FIG. 7 below.

[0065] The prediction device can obtain a second image (530) containing reflection information of a shock wave based on a computed tomography image (510). The prediction device can extract a boundary area of ​​the computed tomography image (510) through image processing and convert the reflection information of the boundary area into attenuation information. A specific method for obtaining the second image (530) is described below with reference to FIG. 8.

[0066] The prediction device can obtain a sinogram (540) including shock wave attenuation information and reflection information based on the first image and the second image. As will be described in detail below, the degree of shock wave attenuation, which is directly related to the success of lithotripsy, can be inferred from the degree of attenuation of the x-ray, and the shock wave reflection information can be obtained from the second image (530).

[0067] The prediction device can obtain the patient's meta information corresponding to the computed tomography image (510), and input the sinogram (540), the first image (520), the second image (5300) and the patient's meta information into the second artificial neural network model, thereby predicting the success or failure of extracorporeal shock wave lithotripsy. The patient's meta information according to one embodiment can include information related to the patient, such as the patient's age, gender, and medical record information. Each of the data described above as the patient's meta information can be input into the second artificial neural network, or one digitized data based on the data described above can be input into the second artificial neural network.

[0068] The first and second artificial neural networks can be trained end-to-end simultaneously, or trained individually. The second artificial neural network model may include a prediction module composed of multiple prediction blocks. The structure of the second artificial neural network model is described below with reference to Figure 9.

[0069] FIG. 6 is a drawing for explaining a sinogram including attenuation information of a shock wave according to one embodiment.

[0070] Referring to Fig. 6, the equation for shock wave attenuation is as follows: Equation 2, where the linear attenuation constant ( ) can be expressed as an exponential relationship.

[0071]

[0072] Mathematical expression 2 is the same form as mathematical expression 3, which is the exponential relationship of the linear attenuation constant (μ) of x-rays when obtaining a computed tomography image.

[0073]

[0074] Comparing mathematical expressions 2 and 3, the only difference is the attenuation constant of the shock wave, which is a physical wave, and the x-ray, which is an electromagnetic wave. If we look at the attenuation constants of the shock wave and the x-ray in each tissue, we can see that although there are differences in the absolute values ​​of attenuation by tissue, the relative attenuation (i.e., the order of the magnitude of the attenuation constant by body tissue) is similar. Therefore, since the attenuation degree of the shock wave, which is directly related to the success of lithotripsy, can be indirectly inferred from the attenuation degree of the x-ray, the prediction device can obtain the sinogram, which is the accumulated attenuation information in the x-ray path, from the computed tomography image through a random transform (Radon Transform).

[0075] FIG. 7 is a drawing for explaining a modified sinogram according to one embodiment.

[0076] Referring to Figure 7, the existing sinogram can be obtained by obtaining the accumulated attenuation information along the path according to the irradiation angle of the X-ray. However, in the case of shock waves in actual lithotripsy, the attenuation information up to the stone in the target kidney is important, and the attenuation information for shock waves that pass the target kidney stone is not necessary.

[0077] Accordingly, a prediction device according to one embodiment can automatically classify target kidney stones from an automatic classification model, and obtain the location of the target kidney in the form of a heat map from the automatic classification result by applying an explainable artificial intelligence technique. Thereafter, the prediction device deletes an image corresponding to an X-ray path behind the target from the automatically obtained target kidney stone location information, thereby removing unnecessary information for inferring accumulated attenuation information, thereby preventing unnecessary attenuation information from being reflected in the sinogram.

[0078] For example, drawing (710) is a computed tomography image including attenuation information for shock waves passing a target kidney stone, and drawing (720) is a sinogram including attenuation information for shock waves passing a target kidney stone.

[0079] On the other hand, the prediction policy according to one embodiment can generate a first image (730) by deleting an image corresponding to an X-ray path behind the target from the target kidney stone location information in a computed tomography image, and can generate a modified sinogram (740) based on the first image (730).

[0080] FIG. 8 is a drawing for explaining a method for obtaining a second image according to one embodiment.

[0081] Referring to FIG. 8, in one embodiment, another prediction device extracts boundaries within a tissue through edge detection image processing to additionally reflect shock wave reflection information before converting into a sinogram, and then converts the degree of reflection in the boundary region into an attenuation degree and displays it in HU units on a CT image. Thereafter, the prediction device can finally obtain a sinogram that takes into account even the reflectivity, and visualize the attenuation information accumulated along the path.

[0082] FIG. 9 is a diagram illustrating a structure of a second artificial neural network according to one embodiment.

[0083] Referring to FIG. 9, a second artificial neural network according to an embodiment may include a prediction module (e.g., RENSE of FIG. 9) composed of a plurality of prediction blocks. The second artificial neural network may sequentially input input data (e.g., a sinogram, the first image, the second image, and meta information of the patient) to a convolution layer, a batch normalization layer, and an activation layer, and input the output thereof to a prediction module. The output of the prediction module may be input to a transition module composed of a 1X1 convolution layer and a pooling layer. Thereafter, the prediction module and the transition module may be repeated, and in the last stage of the second artificial neural network, a global pooling layer and a linearization layer may be passed to output a final output (e.g., whether extracorporeal shock wave lithotripsy was successful).

[0084] The prediction module may be composed of multiple prediction blocks. The prediction block may be composed of a first convolutional layer, an activation layer, a second convolutional layer, and a second activation layer, and the input of the prediction block may be combined with the output of the second convolutional layer and input to the second activation layer.

[0085] The prediction module can determine the output of the prediction module by combining the outputs of each prediction block. For example, the prediction module can be composed of a first prediction block to an n-th prediction block, and the prediction module can determine the final output not only based on the output of the n-th prediction block, but also based on the entire output of each of the first to n-th prediction blocks. In other words, the second artificial neural network model can learn by integrating residual information and previous information to minimize the loss of input information.

[0086] The embodiments described above may be implemented using hardware components, software components, and / or a combination of hardware components and software components. For example, the devices, methods, and components described in the embodiments may be implemented using a general-purpose computer or a special-purpose computer, such as, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate array (FPGA), a programmable logic unit (PLU), a microprocessor, or any other device capable of executing instructions and responding to them. The processing device may execute an operating system (OS) and software applications running on the operating system. Furthermore, the processing device may access, store, manipulate, process, and generate data in response to the execution of the software. For ease of understanding, the processing device is sometimes described as being used alone; however, one of ordinary skill in the art will recognize that the processing device may include multiple processing elements and / or multiple types of processing elements. For example, a processing unit may include multiple processors, or a processor and a controller. Other processing configurations, such as parallel processors, are also possible.

[0087] Software may include computer programs, codes, instructions, or any combination thereof, which may configure a processing device to perform a desired operation or, independently or collectively, command the processing device. The software and / or data may be permanently or temporarily embodied in any type of machine, component, physical device, virtual equipment, computer storage media or devices, or transmitted signal waves, for interpretation by the processing device or for providing instructions or data to the processing device. The software may also be distributed over networked computer systems and stored or executed in a distributed manner. The software and data may be stored on a computer-readable recording medium.

[0088] The method according to the embodiment may be implemented in the form of program commands that can be executed through various computer means and recorded on a computer-readable medium. The computer-readable medium may include program commands, data files, data structures, etc., alone or in combination. The program commands recorded on the medium may be those specially designed and configured for the embodiment or may be those known and available to those skilled in the art of computer software. Examples of the computer-readable recording medium include magnetic media such as hard disks, floppy disks, and magnetic tapes, optical media such as CD-ROMs and DVDs, magneto-optical media such as floptical disks, and hardware devices specially configured to store and execute program commands such as ROMs, RAMs, and flash memories. Examples of program commands include not only machine language codes such as those generated by a compiler, but also high-level language codes that can be executed by a computer using an interpreter, etc.

[0089] Although the embodiments described above have been described with limited drawings, those skilled in the art will appreciate that various technical modifications and variations can be applied based on the above. For example, appropriate results can still be achieved even if the described techniques are performed in a different order than described, and / or components of the described systems, structures, devices, circuits, etc. are combined or combined in a different manner than described, or are replaced or substituted with other components or equivalents.

[0090] Therefore, other implementations, other embodiments, and equivalents to the claims also fall within the scope of the claims described below.

Claims

1. A step of inputting a computed tomography image into a first artificial neural network model to obtain a first image containing target stone information. A step of obtaining a second image including reflection information of a shock wave based on the above computed tomography image; and A step of obtaining a sinogram including attenuation information and reflection information of the shock wave based on the first image and the second image. A method for predicting extracorporeal shock wave lithotripsy, including:

2. In paragraph 1, A step of acquiring patient meta information corresponding to the above computed tomography image A method for predicting extracorporeal shock wave lithotripsy, further comprising:

3. In paragraph 2, A step of predicting the success or failure of extracorporeal shock wave lithotripsy by inputting the above sinogram, the first image, the second image, and the patient's meta-information into a second artificial neural network model. A method for predicting extracorporeal shock wave lithotripsy, further comprising:

4. In paragraph 3, The above second artificial neural network model is It includes a prediction module composed of multiple prediction blocks, The above prediction module A method for predicting extracorporeal shock wave lithotripsy, wherein the output of the prediction module is determined by combining the outputs of each of the plurality of prediction blocks.

5. In paragraph 1, The step of acquiring the above first image is A step of inputting the above computed tomography image into the first artificial neural network model to obtain a target stone location; and A step of generating the first image by modifying the computed tomography image based on the target stone location. A method for predicting extracorporeal shock wave lithotripsy, including:

6. In paragraph 1, The step of acquiring the above first image is A step of inputting the above computed tomography image into the first artificial neural network model to obtain texture information of the target stone. A method for predicting extracorporeal shock wave lithotripsy, including:

7. In paragraph 1, The step of obtaining the above second image is A step of extracting a boundary area of the above computed tomography image; and A step of converting the reflection information of the above boundary area into the attenuation information. A method for predicting extracorporeal shock wave lithotripsy, including:

8. A computer program stored on a medium to execute any one of the methods of claims 1 to 7 in combination with hardware.

9. Input the computed tomography image into the first artificial neural network model to obtain a first image containing target stone information, Based on the above computed tomography image, a second image including reflection information of a shock wave is obtained, Obtaining a sinogram including the attenuation information of the shock wave and the reflection information based on the first image and the second image An extracorporeal shock wave lithotripsy prediction device comprising a processor.

10. In paragraph 9, The above processor An extracorporeal shock wave lithotripsy prediction device that acquires patient meta-information corresponding to the above computed tomography image.

11. In paragraph 10, The above processor An extracorporeal shock wave lithotripsy prediction device that inputs the above sinogram, the first image, the second image, and the patient's meta information into a second artificial neural network model to predict the success or failure of extracorporeal shock wave lithotripsy.

12. In paragraph 11, The above second artificial neural network model is It includes a prediction module composed of multiple prediction blocks, The above prediction module An extracorporeal shock wave lithotripsy prediction device that determines the output of the prediction module by combining the outputs of each of the plurality of prediction blocks.

13. In paragraph 9, The above processor The above computed tomography image is input into the first artificial neural network model to obtain the target stone location, An extracorporeal shock wave lithotripsy prediction device that generates the first image by modifying the computed tomography image based on the target stone location.

14. In paragraph 9, The above processor An extracorporeal shock wave lithotripsy prediction device that inputs the above computed tomography image into the first artificial neural network model to obtain texture information of a target stone.

15. In paragraph 9, The above processor Extracting the boundary area of the above computed tomography image, An extracorporeal shock wave lithotripsy prediction device that converts the reflection information of the above boundary area into the attenuation information.

Images